ASTR is an indigenous and innovative SIM fraud detection tool, using facial recognition technique powered by Artificial Intelligence and Machine Learning for carrying out proactive analysis, generating communication intelligence for identifying non-bonafide SIMs and weeding them out from the telecom ecosystem before they are used for carrying out cyber fraud or any other crimes. It shares a list of fraudulent connections with telecom operators to disconnect them and shares with LEAs for suitable action as per the rule of law. ASTR also assists LEAs by providing them the list of those Point Of Sale (PoS) which get involved with fraudsters and issue illegal SIM cards. The vision of solution is to ensure a robust and trustworthy telecommunication ecosystem to keep the fake SIMs and fraudsters away from the telecom network.
C-DOT | ASTR (AI and Facial Recognition powered Solution for Telecom SIM Subscriber Verification)
- Scale & Scalability:
- It is the first ever large-scale Pan-India SIM subscriber base image analysis
- Effectively uses state of the art AI/ML technologies
- Highly scalable solution to handle the scale (~140cr facial data) and divergence of faces in India.
- Feature Representation:
- Rich feature representation such that all SIM subscribers PAN India can be easily distinguishable from each other whereas varying images of same subscriber can still have similar representations
- Takes care of the tilt and angle of the face, varying facial skin tones, various India specific ethnic wears and head gears (turbans, gamchha, chunni, burqa etc.) which occlude the face
- Searching & Clustering: A horizontally scaled, distributed, searching algorithm has been developed for large-scale facial comparison.
- High Precision and High recall
ASTR has its use cases in any domain where there is facial verification. Any database of faces can be transformed into a vector database and subsequently self-similarity or query based facial search can be conducted.
ASTR is already been deployed in the following domains-
- Telecom SIM subscriber database de-duplication
- 134 crore mobile connections already analyzed and detected more than 67 lakh fraudulent connections out of which 59 lakh connections have been disconnected by TSPs after verification.
- Dead body identification in accidents
- In Balasore train accident in June 2023, ASTR helped in identifying around 165 unidentified dead bodies.
- Criminal search and identification by Law Enforcement Agencies (LEAs)
- Missing person identification
- Email: support@example.com
- Phone: +91-9876543210
- Address: 123, Tech Park, Bangalore, India